Semantic Segmentation Model for Transmission Tower Point Cloud Based on Improved PointNet++

doi:10.11930/j.issn.1004-9649.202206087

Abstract

Abstract: Aiming at the existing problem that the point cloud extraction accuracy of transmission lines is not high and cannot meet the needs of autonomous and refined inspection by unmanned aerial vehicles, an improved PointNet++ semantic segmentation method for transmission tower point cloud is proposed, which realizes the segmentation of wires, ground wires, drainage lines, insulators and towers. Firstly, the parameters of the classic PointNet++ model are adjusted to make the model more suitable for the point cloud data of transmission towers in terms of feature extraction quantity and receptive field; then, the kernel point convolution is used as the point cloud feature extraction algorithm to further improve the model's ability to detect point cloud features; finally, for the data imbalance problem in the point cloud data, the focal loss is used as the loss function, so that the categories with a small proportion can be fully trained. In order to verify the effectiveness of the proposed method, experiments are carried out on the point cloud dataset composed of 2284 transmission towers. The experimental results show that the average F₁ value of the improved algorithm reaches 97.26%, which is 3.95 percentage points higher than that of the classic PointNet++.

Key words: transmission tower, point cloud segmentation, kernel point convolution, focal loss, PointNet++

HUANG Zheng, GU Xu, WANG Hongxing, ZHANG Xingwei, ZHANG Xin. Semantic Segmentation Model for Transmission Tower Point Cloud Based on Improved PointNet++[J]. Electric Power, 2023, 56(3): 77-85.

Add to citation manager EndNote|Ris|BibTeX

URL: https://www.electricpower.com.cn/EN/10.11930/j.issn.1004-9649.202206087

https://www.electricpower.com.cn/EN/Y2023/V56/I3/77

References

[1] 邵瑰玮, 刘壮, 付晶, 等. 架空输电线路无人机巡检技术研究进展[J]. 高电压技术, 2020, 46(1): 14–22
SHAO Guiwei, LIU Zhuang, FU Jing, et al. Research progress in unmanned aerial vehicle inspection technology on overhead transmission lines[J]. High Voltage Engineering, 2020, 46(1): 14–22
[2] 邓恢平. 无人机在输电线路应用不断深化: 从辅助支撑作用到巡检主力军[N]. 中国电力报, 2021-12-23(8).
[3] 黄郑, 王红星, 周航, 等. 基于混合算法的电力杆塔巡检实时航迹规划[J]. 中国电力, 2021, 54(11): 214–220
HUANG Zheng, WANG Hongxing, ZHOU Hang, et al. Real-time path planning for power tower inspection based on hybrid algorithm[J]. Electric Power, 2021, 54(11): 214–220
[4] 黄郑, 王永强, 王红星, 等. 基于云雾边异构协同的无人机智慧巡检系统[J]. 中国电力, 2020, 53(4): 161–168
HUANG Zheng, WANG Yongqiang, WANG Hongxing, et al. Design and application of UAV intelligent inspection system for transmission lines based on cloud and fog-edge heterogeneous collaborative computing architecture[J]. Electric Power, 2020, 53(4): 161–168
[5] 麻卫峰, 王成, 王金亮, 等. 基于激光点云的高压输电线覆冰厚度反演[J]. 电力系统保护与控制, 2021, 49(4): 89–95
MA Weifeng, WANG Cheng, WANG Jinliang, et al. Inversion of ice thickness for high voltage transmission line based on a LiDAR point cloud[J]. Power System Protection and Control, 2021, 49(4): 89–95
[6] 廖如超, 张英, 廖建东, 等. 基于语义信息分块的高像素导线缺陷目标识别[J]. 电力科学与技术学报, 2022, 37(3): 206–212
LIAO Ruchao, ZHANG Ying, LIAO Jiandong, et al. Research on defect target identification of high pixel wire image based on semantic information patching[J]. Journal of Electric Power Science and Technology, 2022, 37(3): 206–212
[7] 陈超强, 龚汉阳, 张帝, 等. 计及设备信息的配电设备多组巡检路径优化策略[J]. 中国电力, 2022, 55(7): 81–86
CHEN Chaoqiang, GONG Hanyang, ZHANG Di, et al. Optimization strategy of multi-group inspection path of distribution equipment with equipment information included[J]. Electric Power, 2022, 55(7): 81–86
[8] 周汝琴, 许志海, 彭炽刚, 等. 一种高压输电走廊机载激光点云分类方法[J]. 测绘科学, 2019, 44(3): 21–27, 33
ZHOU Ruqin, XU Zhihai, PENG Chigang, et al. A JointBoost-based classification method of high voltage transmission corridor from airborne LiDAR point cloud[J]. Science of Surveying and Mapping, 2019, 44(3): 21–27, 33
[9] 王和平, 陈世超, 胡伟, 等. 机载LiDAR输电线路杆塔快速定位方法研究[J]. 遥感技术与应用, 2021, 36(6): 1306–1310
WANG Heping, CHEN Shichao, HU Wei, et al. Study on power pylon fast positioning in transmission line from airborne LiDAR data[J]. Remote Sensing Technology and Application, 2021, 36(6): 1306–1310
[10] 徐梁刚, 虢韬, 吴绍华, 等. 基于点云数据特征的电力线快速提取和重建[J]. 激光技术, 2020, 44(2): 244–249
XU Lianggang, GUO Tao, WU Shaohua, et al. Fast extraction and reconstruction of power line based on point cloud data features[J]. Laser Technology, 2020, 44(2): 244–249
[11] 谭弘武, 王敬茹, 刘武能, 等. 机载LiDAR高压线塔点云自动化提取[J]. 遥感信息, 2021, 36(4): 7–11
TAN Hongwu, WANG Jingru, LIU Wuneng, et al. Automatic extraction of high voltage line tower point cloud from airborne LiDAR data[J]. Remote Sensing Information, 2021, 36(4): 7–11
[12] 朱依民, 田林亚, 毕继鑫, 等. 基于无人机机载LiDAR的电力线点云提取与重建[J]. 激光技术, 2021, 45(5): 554–560
ZHU Yimin, TIAN Linya, BI Jixin, et al. Power line point cloud extraction and reconstruction based on UAV-borne LiDAR[J]. Laser Technology, 2021, 45(5): 554–560
[13] 曾远, 陈亦, 姚攀, 等. 基于机载激光点云数据的输电线走廊自动识别技术[J]. 应用激光, 2021, 41(5): 1033–1038
ZENG Yuan, CHEN Yi, YAO Pan, et al. Automatic identification of transmission line corridors based on airbornelaser point cloud data[J]. Applied Laser, 2021, 41(5): 1033–1038
[14] 彭淑雯. 输电通道机载LiDAR点云分类方法研究[D]. 北京: 中国科学院大学(中国科学院空天信息创新研究院), 2021.
PENG Shuwen. Research on airborne LiDAR point cloud classification method for transmission line[D]. Beijing: Aerospace Information Research Institute, Chinese Academy of Sciences, 2021.
[15] 陈正宇, 彭淑雯, 朱号东, 等. 基于样本加权PointNet++的输电通道点云分类研究[J]. 遥感技术与应用, 2021, 36(6): 1299–1305
CHEN Zhengyu, PENG Shuwen, ZHU Haodong, et al. LiDAR point cloud classification of transmission corridor based on sample weighted-PointNet[J]. Remote Sensing Technology and Application, 2021, 36(6): 1299–1305
[16] ELDAR Y, LINDENBAUM M, PORAT M, et al. The farthest point strategy for progressive image sampling[J]. IEEE Transactions on Image Processing, 1997, 6(9): 1305–1315.
[17] ARMENI I, SENER O, ZAMIR A R, et al. 3D semantic parsing of large-scale indoor spaces[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, NV, USA. IEEE, 2016: 1534–1543.
[18] DAI A, CHANG A X, SAVVA M, et al. ScanNet: richly-annotated 3D reconstructions of indoor scenes[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI, USA. IEEE, 2017: 2432–2443.
[19] HU Q Y, YANG B, XIE L H, et al. RandLA-net: efficient semantic segmentation of large-scale point clouds[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle, WA, USA. IEEE, 2020: 11105–11114.
[20] HUANG G, LIU Z, VAN DER MAATEN L, et al. Densely connected convolutional networks[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI, USA. IEEE, 2017: 2261–2269.
[21] HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, NV, USA. IEEE, 2016: 770–778.
[22] THOMAS H, QI C R, DESCHAUD J E, et al. KPConv: flexible and deformable convolution for point clouds[C]//2019 IEEE/CVF International Conference on Computer Vision (ICCV). Seoul, Korea (South). IEEE, 2020: 6410–6419.
[23] CHARLES R Q, HAO S, MO K C, et al. PointNet: deep learning on point sets for 3D classification and segmentation[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI, USA. IEEE, 2017: 77–85.
[24] QI C R, YI L, SU H, et al. PointNet++: deep hierarchical feature learning on point sets in a metric space[C]//Proceedings of the 31 st International Conference on Neural Information Processing Systems. Long Beach, California, USA. New York: ACM, 2017: 5105–5114.
[25] LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[C]//IEEE Transactions on Pattern Analysis and Machine Intelligence. IEEE, 2018: 318–327.